Guide 1 · Reading prices
How to read prediction-market prices as probabilities
A share that pays $1 if something happens and trades at 58¢ implies roughly a 58% chance. That reading is a useful shortcut, but what you actually pay also includes fees, spreads and price impact, and the price is the market’s estimate, not a fact.
From cents to a chance
A YES share pays exactly $1 if the event happens and $0 if it doesn’t. If you think the chance is 70%, the fair price for you is 70¢: over many such bets you would break even. When many people trade, the price settles near the level where buyers and sellers balance, so 58¢ reads as “the market thinks about 58%”. Move the slider:
58¢ YES at 58¢ means the market prices about a 58% chance. NO trades near 42¢.
Why YES and NO add up to about $1
Exactly one of YES and NO will pay $1. So holding one of each is worth exactly $1 at settlement, and in theory their prices sum to $1. On 369X’s LMSR market maker, the quoted prices do sum to 1. What you pay to buy is higher, because of the fee and the price impact of your own order. The same is true of selling in reverse.
What you actually pay
Hypothetical Buying $100 of YES quoted at 40¢
- Fee
- 2% of $100 = $2
- Price impact
- Each share you buy nudges the price up, so your average might be 40.4¢, not 40¢
- Effective breakeven
- At a 40¢ average after the fee, you need the event to happen about 41% of the time to break even on average
The payout calculator on the fees page works through these numbers for any spend and average price.
Spreads, and why quiet markets mislead
On an order book, the best buy price and the best sell price differ; that gap is the spread. A market showing “YES 60¢” may really be “sellers at 64¢, buyers at 56¢”, and the midpoint hides the cost of getting in and out. On an LMSR market there is no spread in the same sense, but a thin market has a small liquidity parameter, so even modest trades move the price a lot.
In both cases, a price in a quiet market is weaker evidence than the same price in a busy one.
What moves a price
- News. An injury, a poll, a statement. The fastest moves happen in the minutes after information breaks.
- Order flow. A large buyer moves the price whether or not they know anything.
- Time. As the end date nears with nothing happening, prices on “will X happen by date Y” drift toward NO.
- Hedging. Someone offsetting a risk elsewhere may trade against their own belief.
What a good price looks like over time
A single market can’t prove a price was right or wrong. A 70% favourite that loses isn’t a failure; it should happen three times in ten. The test is calibration across many markets: of all the markets priced near 70%, about 70% should resolve YES. That is how forecasts, including polls and markets, are properly judged. Markets versus polls.
In short
Price ≈ probability is a good first read. Then adjust for fees, price impact, how busy the market is and how close the end date is.